Testing treatments

Boots the chemist are making much of research just published in the British Journal of Dermatology. It involves their Protect and Perfect product, which is on sale in my local store, where there are signs saying that customers are allowed to buy only 6 bottles. Clearly they are anticipating great demand.

In 2007, the BBC program Horizon revealed results of a test of Protect and Perfect. They used it on 15 people over 12 days, applying the cream to the forearm and then assaying skin samples for fibrillin, which – say Boots may ”perhaps clinically improve photoaged skin”. 

Rightly, this experiment was criticised, and it was announced that proper trials were going to be done, and here, in the BJD, they are.

This is what Boots says about it:

“The trial was an independent study of 60 volunteers over an initial period of six months. Half the volunteers used No7 Protect & Perfect Intense Beauty Serum and the other half used a placebo product (the same formula, but without the anti-ageing ingredients). Neither group knew which product they had been given.

The results for No7 Protect & Perfect Intense Beauty Serum were astounding. After the initial six months were over, skin on the volunteers using the real No7 Protect & Perfect Intense Beauty Serum showed some repair of the damage caused by sun exposure. No changes were seen for the group using the placebo product.

Use of the No7 Protect & Perfect Intense Beauty Serum for an additional six months led to clinical reduction in the appearance of wrinkles, assessed by an independent dermatologist. Seventy per cent of people perceived a marked improvement in their skin, thus proving that the skin looks better the longer you use No7 Protect & Perfect Intense.”

I don’t think that’s quite how I’d put it. Certainly, it was an RCT, and that’s good. The 60 participants were blinded to either the placebo cream or the test cream for six months. After six months the groups were unblinded and they all used the test cream. However the study used linear regression analysis to “extrapolate the vehicle response to 12 months, thus allowing comparisons with the test product”. This seems like a bad idea to me: this technique is not as true to life as doing the comparative test in real life would be.  The assessments of skin changes were made by two dermatologists, and I am not convinced that we know this is a reliable and reproducible tool. But very interestingly, in the abstract summary it says “at six months, the test product produced statistically significant improvement in facial wrinkles as compared to baseline assessment (p=0.013) whereas vehicle-treated skin was not significantly improved (p=0.11).” A p value of less than 0.05 is usually treated as being statistically significant in medical papers. However, in the discussion section, please note (and my caps):

“Compared to the baseline, the test product did lead to a noticeable clinical improvement in facial wrinkles (P=0.013) in 43 per cent of treated individuals after six months, compared with only 22 per cent of those treated with the vehicle where there was no significant improvement in appearance (p=0.11). In a comparison between groups THIS IMPROVEMENT WAS NOT STATISTICALLY SIGNIFICANT but does indicate that larger clinical trials of cosmetic products might be expected to show useful clinical improvement after six months’ use.”

I interpret that as bit of a wish-list. I’m not astounded either.

Or perhaps not so much a makeover, but a radical shift in how drug research is decided upon, performed and reported. The suggestions come from Sir Iain Chalmers, who is editor of the James Lind Library in Oxford, and Silvio Garattini, director of the Mario Negri Institute for Pharmacological Research in Milan. Writing in the British Medical Journal, they say that their proposals would not only benefit patients but also industry, which is not winning any prizes for popularity at present.

They are:

1) Patients to be involved in shaping the research agenda – in other words, making sure research questions have resonance in real life and real-world situations

2) Legal requirements for research to be published, including trial protocols, by all (mandatory publication of trial data has had legal backing in the US since late 2007; no such protection for patients currently exists in the UK)

3) Independent evaluation of drugs. As the paper says: “The monopoly that the drugs industry has in evaluating its own products, and the secrecy surrounding this process, leads to biased evidence that is currently only rarely questioned by independent studies.”

4) A requirement to demonstrate “added value” for all new drugs – is this drug better than the current best drug treatment, or does it benefit in addition to it, and is it better than non-drug treatments? Too often, trials are done comparing a new treatment to placebo where there is a known intervention which is better than placebo. This means that uncertainty about how to use it best persists.

The authors say that these would help improve public confidence in pharma, would improve returns from investment in R&D, but could also improve efficiency in other ways. For example: “Tim Mant, a director of a major contract research organisation, has acknowledged how frustrating it is to be commissioned to organise a clinical trial that he knows is going up a scientific blind ally because he has been there previously with another company but cannot divulge information that is commercially confidential.”

I hope that people in the pharmaceutical industry take notice: I would genuinely love to write about the improvements that are being made.

I was talking to a composer a few weeks ago. “This stuff doesn’t really exist except when it’s played,” he said, pointing to his score with heavy despair. “Whereas you’ve got a job where you can actually see that you are doing something good.” He couldn’t understand that my protests to the contrary were genuine: doctors cannot always be sure that they are making a positive difference.

Medical history is stuffed with examples of bad practice – lobotomy for just about anyone with a mental health problem, tonsillectomy for most people with a sore throat, bed rest for everyone with low back pain. I may exaggerate, but only a little. Even now, we don’t seem to appreciate the value of analysing what we are doing.

The US Senate has just passed a bill putting $1.1bn into research aimed at identifying which medical treatments work and which don’t. This sounds like a good idea, but of course not all research is created equal. Regular readers will be aware, for example, of the conflicts surrounding the Department of Health recommendation that all over-65s receive a flu vaccination. Some evidence supports vaccination, some indicates that it does not improve mortality rates. What to believe?

The remainder of the article can be read here. Please post comments below.

Invitations have been pouring through letterboxes all over the UK to take part in the Biobank.

If you haven’t received one already, let me explain what this particular bank wants from you (thankfully, it doesn’t involve money). The Biobank is a research project, and its aim is no less than to improve the “health of future generations”.

Funded by the Medical Research Council, the Wellcome Trust and the Department of Health, among other bodies, it is recruiting half a million people between 40 and 69 to be surveyed about their health. They will be followed for several decades, in some cases until death. People will be asked about health, lifestyle, work, family history, and have blood and urine samples taken for storage. They will also have tests for blood pressure, bone density and lung function. The researchers may ask for permission to access medical notes, and they may in future examine blood for genetic factors.

The remainder of the article can be read here. Please post comments below.

Sir Richard recently gave an interview to the BBC  when he said, amongst other things, that the healthcare industry could learn from the airline industry; and that all healthcare workers should be screened for MRSA and treated for it because it “is far better than having people dying from unnecessary diseases, and all the misery and pain that that causes, and the cost to the NHS which is enormous.”

Sir Richard is now vice-chair of the Patients Association. If he wanted to go and talk to the scientists who actually do know about MRSA then he would find out all kinds of things; for example, in many outbreaks of MRSA, staff strains are different from those that patients are colonised by. And that MRSA is on places that may not routinely get cleaned; and that it is a bit daft to be so concerned about cleaning bedposts if there is only one commode being shared by a whole ward. Now, if Sir Richard was proposing research to find out what the most cost-effective ways are of reducing MRSA (and other hospital aquired pathogens) transmission and disease resulting from it are, I would be entirely supportive. But presuming that one knows the answers when it is clear that this is a complex area where randomised controlled trials are few – is dangerous.

As for the airline/healthcare analogy, well…

If a pilot thinks it’s unsafe to fly due to risk factors, for example poor weather, then they don’t. They stay, rightly, grounded. If a doctor thinks that surgery will be high risk, they don’t always have the choice of staying ‘grounded’ and not operating: the illness may well be the reason why the operation needs to be done. In other words, the airline industry has much more choice about the risks it is prepared to take on.

And. Airlines fly routes that are profitable and readily possible. Healthcare has to deal with things that may be neither. Neither can the identification of ‘near misses’ in air travel be used as a reason to compare it with safety in healthcare – in any case there seems to be justified concern that pilots don’t always ‘fess up.

This isn’t to diminish the huge responsibility which airline pilots take on and have. Aviaton and healthcare systems may have some similarities but they are limited. Here is one comparision it might be worth making. A pilot has a co-pilot and a standard number of crew without whom he cannot fly. The healthcare vogue is for promoting less qualified team members to diagnose and treat conditions. This is analogous to the pilot remaining at the airport but taking responsibility for the cabin crew flying the aircraft and dealing with any problems. It may be cheaper to do so but it isn’t necessarily desirable or effective. This is something which competitors to NHS primary healthcare may wish to note.

The only thing separating reiki and reflexology from rational medicine and progress is evidence. And what evidence-based aficionados like me love best is the randomised controlled trial, the process by which most drugs and treatments aspire to be tested. But are we now hearing its death rattle?

In a recent lecture at the Royal College of Physicians, Professor Sir Michael Rawlins appeared to have stuck the knife in the randomised controlled trial. Sir Michael, chairman of the National Institute for Health and Clinical Excellence, effectively said that such trials weren’t the be-all and end-all. “Sir Michael Rawlins attacks traditional ways of assessing evidence”, the story went. But it is not the value of these trials that is the problem, but rather how we chose to think about their conclusions.

The remainder of the article can be read here. Please post comments below.

Much ado with a new paper published by the New England Journal of Medicine . This study was placebo controlled and focused on treating people with ”normal” cholesterol but a high “c-reactive protein” (a marker of inflammation) with rosuvastatin (which is not a new statin as some media outlets have reported, but one already in use). Reports have been rather enthusiastic, eg from the Daily Telegraph: ” risk of a heart attack was reduced by 54%”. One doctor is reported as saying it’s “astonishing”. The trial was stopped early due to “remarkable” results.

The problem is that although the “54%” looks marvellous, and is true, this is the relative risk reduction, not the absolute risk reduction. It does not, by itself, give us a true picture of how meaningful this reduction in cardiovascular events is. We have to know what our risk of having such an event was to start with. From “table 3″ in the paper, the number of patients in the rosuvastatin group was 8901. The number of heart attacks in this group was 31. The placebo group was also made up of 8901 people. The number of heart attacks in the placebo group was 68.  The chance of this group of people having a heart attack on placebo treatment was 68/8901, or 0.76%. The chance of the other group of people, those on rosuvastatin, having a heart attack were 31/8901, or 0.35%. Thus, if you have a normal cholesterol but a high CRP, and if you take rosuvastatin, you can have a 0.35% chance of having a heart attack as opposed to a 0.76% chance.

I’m not very impressed. The other problem with this trial is that it was stopped early. Thus we don’t know what the long term benefits or problems of this approach were (article on this here) . And there did seem to be a small increased risk of developing diabetes in the rosuvastatin group.

However, there may be something else going on here. I mentioned the thought-provoking book The Cholesterol Con by Dr Malcolm Kendrick a while ago. He says, effectively, that cholesterol is nothing to do with heart disease. Statins seem to have some effect on outcomes, but probably have another way of working which has nothing to do with cholesterol, but something to do with inflammation.

There is an interesting study this week in the BMJ. The study was a mailed survey to US internists and rheumatologists about their use of placebo treatments. The response rate wasn’t great (57%) but about half said they regularly prescribed placebo treatments. Most also said they thought it was ethically permissible.

Placebos do work and the placebo response is usually a clinically useful one. The question is how to use it practically without deceiving the patient. (I am not aware of any research that explores how the placebo effect varies according to what the patient is told about what the treatment contains; do let me know if you do know of any.)

Ethically, doctors should not deceive by lying or exaggerating what is being given. Some ethicists have postulated that by giving a placebo treatment and saying something like ‘we don’t really know how the treatment we are going to give you will work, but I believe it will, and it will not cause any side effects’ is okay. My  problem is that I am not quite sure this is a good enough explanation; I find the explicit omission uncomfortable.

However, the ‘placebo effect’ can be very usefully and ethically harnessed by way of ‘placebo-like’ effects,  the effect more generally of an ongoing relationship between patient and doctor. For example, continuous care from the same doctor, longer appointment times and empathy, all result in better outcomes for patients . The political direction that primary care has been sent in, though, hardly allows for the importance of these things to flourish.

Chronic obstructive pulmonary disease is a blight on British health, estimated to cause 20 per cent of medical hospital admissions. Primarily caused by cigarette smoking, it is a condition that damages the airways and obstructs the flow of air from the lungs, leading to breathlessness, a chronic cough and wheezing. The symptoms are distressing and, as one might imagine, they have a significant effect on quality of life.

So what treatments for COPD can improve a patient’s wellbeing, mood and sense of control over the condition? The answer may be surprising, in part because we have become accustomed to hearing almost daily about “breakthroughs” in genetic decoding and state-of-the-art biotechnology. New drugs are permanently “on the horizon”.

 The rest of this column can be read here. Please post comments below.

The more one learns, the less one is certain of. This is as true in medicine as in life. Thousands of clinical studies have investigated the effects of hypertension treatments, yet there is still considerable uncertainty about which drugs are most effective.

Meanwhile, many people with no history of cardiovascular disease are being diagnosed with hypertension, and encouraged to take medication for it. But what works best?

The Drugs and Therapeutics Bulletin, a slender but valuable journal, recently published a review of the current evidence on hypertension. There are four main groups of drugs to treat high blood pressure. The first, bendrofluazide, is a diuretic, meaning that urine production is increased, which makes it unpopular with some. Beta-blockers, such as atenolol, are another choice; but these are unsuitable for people with asthma and often trigger side effects, including impotence. Calcium channel blockers such as amlodipine are useful agents but can cause swollen ankles. And ACE inhibitors, such as enalapril, can cause renal problems and require monitoring.

The rest of this column can be read here. Please post comments below.

Margaret McCartney’s Blog

This blog is no longer updated but it remains open as an archive.

A forum on healthcare policy and professional issues, by Glasgow-based GP and FT Weekend columnist Margaret McCartney.

FT Blogs